Epidemiology and temporal trendof suicide mortality in the elderlyin Jiading, Shanghai, 2003–2013:a descriptive, observational study

Yueqin Shao,1 Chenghua Zhu,2 Yiying Zhang,1 Hongjie Yu,1 Hui Peng,1

Yaqing Jin,1 Guozheng Shi,1 Na Wang,2 Zheng Chen,1 Yue Chen,3 Qingwu Jiang2

To cite: Shao Y, Zhu C,Zhang Y, et al. Epidemiologyand temporal trend of suicidemortality in the elderlyin Jiading, Shanghai, 2003–2013: a descriptive,observational study. BMJOpen 2016;6:e012227.doi:10.1136/bmjopen-2016-012227

▸ Prepublication history forthis paper is available online.To view these files pleasevisit the journal online(http://dx.doi.org/10.1136/bmjopen-2016-012227).

YS and CZ contributedequally.

Received 9 April 2016Revised 17 July 2016Accepted 19 July 2016

For numbered affiliations seeend of article.

Correspondence toDr Na Wang; na.wang@fudan.edu.cn

ABSTRACTObjectives: To investigate and describe theepidemiological characteristics of suicide in the elderlyin Jiading, Shanghai, for the period 2003–2013.Design: Retrospective, observational, epidemiologicalstudy using routinely collected data.Setting: Jiading District, Shanghai.Methods: Suicide data were retrieved from theShanghai Vital Registry database for the period 2003–2013. Crude and age-standardised mortality rates werecalculated for various groups according to sex andage. Joinpoint regression was performed to estimatethe percentage change (PC) and annual percentagechange (APC) for suicide mortality.Result: A total of 956 deaths due to suicide occurredamong people aged ≥65 years during the study period,accounting for 76.7% (956/1247) of all suicidedecedents. Among the 956 people with suicide deaths,88.7% (848/956) had a history of a psychiatriccondition. The age-standardised mortality rates forsuicide without and with a psychotic history in peopleaged ≥65 years were much higher than those forpeople aged <65 years in both genders. Suicidemortality in the elderly showed a declining trend, witha PC of −51.5% for men and −47.5% for women.The APC was −29.1 in 2003–2005, 4.6 in 2005–2008and −9.7 in 2008–2013 for aged men, and −12.2 in2003–2006 and −5.2 in 2006–2013 for aged women,respectively. Women living in Jiading had a higher riskof suicide death than men, especially among theelderly. The mortality rate for suicide increased withage in the elderly, and was more marked for those witha psychiatric history than for those without.Conclusions: Suicide mortality declined in Jiadingduring the study period 2003–2013 overall, butremained high in the elderly, especially those with apsychiatric history.

INTRODUCTIONIn 2012, 8.04 million people died fromsuicide, accounting for 1.4% of the totaldeaths worldwide, and suicide was the 15th

leading cause of death.1 According to datafrom the Chinese Center for Disease Controland Prevention, suicide was the secondleading cause of injury death for the generalpopulation, while death from all types ofinjury was the fourth most common cause ofdeath after cardiovascular diseases, cancerand respiratory diseases.2

Suicide is a serious public health problemfor the elderly, who constitute the greatestdemographic group and have the highestsuicide mortality in most countries.1 In devel-oped countries, elderly men are at higherrisk than elderly women,3 but the gender dis-parity is less pronounced in developing coun-tries.4 Psychiatric illness is present in mostsuicide deaths in the elderly (71–97%), withaffective disorders being the most common.3

In the past two decades, overall suicidemortality has decreased remarkably in China,which might partly be due to the improve-ment in healthcare system and economicdevelopment.5 Suicide mortality shows somechanging patterns, including a reduced dis-crepancy between urban and rural residentsand an increased male/female ratio.6 Theimpact of psychiatric history on suicide

Strengths and limitations of this study

▪ This study describes the epidemiology of suicidemortality for elderly people with and without apsychiatric history. To date, there has beenlimited research into the effect of psychiatrichistory on suicide death in the elderly.

▪ The data were retrieved from the Shanghai VitalRegistry database and focused on the populationfrom a suburban area in the process ofurbanisation.

▪ The main limitations of the study are the lack ofdetailed information on the type and severity ofpsychiatric disease and information aboutsuicide attempts.

Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227 1

Open Access Research

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from

mortality and gender difference has not been well inves-tigated in the elderly, and the suicide mortality rate databased on complete surveillance are limited.5 With popu-lations ageing and rapid urbanisation together with life-style changes, suicide mortality in the elderly is animportant public health problem in China. This studyexamines the temporal trend and epidemiologicalcharacteristics of suicide mortality, especially among theelderly.

METHODS AND MATERIALSDataJiading is a suburban district located in northwestShanghai, with a revenue of 66 billion Chinese yuan in2014. Of the total population of 5.76 million in 2014,4.94 million were living in urban areas and the male/female ratio was 0.99. In this study, we used dataextracted from the Shanghai Jiading Vital Registry data-base, which were provided by the Jiading District Centerfor Disease Control and Prevention, Shanghai. Thecauses of death were coded using the 10th edition ofthe International Classification of Diseases (ICD-10).Deaths for suicide were coded under external causes ofmorbidity, labelled V00-Y99.7 Suicide decedents were allregistered permanent residents whose underlying causeof death was ‘suicide’. Information was also obtainedabout comorbidities, including psychiatric disorders,which had been previously diagnosed by physicians andrecorded on death certificates. All cases of suicide deathwere then divided into two groups: suicide deaths inpeople without and with a psychiatric history. Weobtained demographical data of the general populationfrom the Jiading Public Security Department. All privateinformation including ID was blinded before conductingdata analysis to meet the requirements of ethic state-ments. The study was approved by the ethical reviewboard of the School of Public Health of Fudan University.

Data analysisThe study population was separated into two age groups(<65 and ≥65 years). For each of these two groups, wecalculated the crude and age-standardised mortalityrates for external causes, suicide and suicide withoutand with psychiatric history, respectively. To describechange over time, the percentage change (PC) was cal-culated using the following formula:

PC% ¼ ðrx þ rx�1Þ � ðr1 þ r2Þðr1 þ r2Þ � 100%

where r1 and r2 are the first and second year’sage-standardised mortality rates; rx and rx−1 are the lastand next to last year’s age-standardised mortality rates,respectively.We also analysed change over time using Joinpoint

Trend Analysis Software, which was initially used by the

Table

1Age-standardisedmortalityrate

forsuicide(per100000)in

Jiading,2003–2013

Suicide

Suicidewithoutpsychiatric

history

Suicidewithpsychiatric

history

Male

Female

Male

Female

Male

Female

Age(years)

Total

<65

≥65

Total

<65

≥65

Total

<65

≥65

Total

<65

≥65

Total

<65

≥65

Total

<65

≥65

2003

14.22

5.81

126.00

16.82

6.85

149.26

3.99

3.66

8.30

5.28

4.52

15.3

10.24

2.15

117.7

11.55

2.33

133.96

2004

8.90

4.24

70.76

11.75

4.90

102.78

4.66

3.45

20.66

4.12

3.51

12.23

4.24

0.79

50.11

7.63

1.39

90.55

2005

9.16

4.53

70.70

13.10

5.66

111.89

2.85

2.20

11.54

4.42

4.25

6.80

6.31

2.33

59.15

8.67

1.41

105.08

2006

7.89

4.56

52.18

8.14

2.06

88.91

4.09

4.27

1.65

2.41

1.86

9.75

3.80

0.28

50.52

5.73

0.20

79.16

2007

7.15

2.99

62.44

8.03

2.44

82.25

2.86

2.44

8.47

2.02

1.73

5.91

4.30

0.56

53.97

6.01

0.72

76.34

2008

9.70

4.23

82.43

8.21

2.03

90.32

4.69

3.71

17.74

1.99

1.61

7.10

5.01

0.51

64.68

6.22

0.42

83.22

2009

7.00

3.12

58.57

8.48

3.98

68.32

2.62

2.37

5.97

2.79

2.46

7.08

4.38

0.75

52.60

5.70

1.52

61.24

2010

6.85

4.46

38.68

8.54

3.55

74.80

4.25

3.70

11.57

3.12

2.18

15.54

2.60

0.76

27.11

5.42

1.37

59.26

2011

6.48

3.79

42.22

8.24

4.28

60.94

3.34

3.21

5.07

3.95

3.39

11.33

3.14

0.58

37.15

4.29

0.88

49.61

2012

6.04

2.20

57.00

7.41

2.45

73.38

1.69

1.29

7.01

3.08

2.17

15.21

4.35

0.91

49.99

4.33

0.28

58.16

2013

8.39

6.12

38.45

6.74

2.81

59.05

4.39

4.32

5.42

2.79

2.32

8.98

3.00

1.81

33.03

3.96

0.48

50.07

2 Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227

Open Access

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from

National Cancer Institute to analyse cancer trends andhas been widely used in research into suicide.8 9

Briefly, the joinpoint regression analysis is fitted to alog-linear regression model and uses a Monte Carlo per-mutation method to test whether an apparent change intrend is statistically significant.8 9 The analysis starts withthe minimum number of joinpoints (usually 0 joinpoint,which is a straight line) and tests whether one or morejoinpoints are statistically significant and should beadded to the model (up to the maximum number sup-plied by the user). Finally, age-standardised mortalityrate, annual percentage change (APC) and its 95% CIare estimated. The model can be expressed as follows:

lnðrateÞ ¼ aþ bðyearÞ þ 1

where α is a constant term; β a regression coefficientand ε an error term. APC can then be calculated asfollows:

APC% ¼ 100� ðeb � 1Þ

APC is regarded as statistically significant when its 95%CI does not include zero. The joinpoint regressionanalysis was applied to raw mortality data according tosex, age and cause of death. Age-standardised mortalityrates were calculated using the Segi World StandardPopulation.10

Excel and SAS, V.9.3 (SAS Institute, Inc, Cary, NorthCarolina, USA) were used for analysis and all tests of stat-istical significance were based on two-sided probability.

RESULTSFrom 2003 to 2013, a total of 1247 deaths of suicidewere identified, accounting for 26.9% (1247/4639) ofall-cause deaths and 40.3% (1247/3091) of deaths fromexternal causes. Among all suicide decedents, peopleaged ≥65 years accounted for 76.7% (956/1247).Among the 956 people with suicide deaths, 88.7% (848/956) had a history of a psychiatric condition.The average mortality for suicide during 2003–2013

was 21.04/100 000. The average mortality rate forsuicide was 4.30/100 000 for those aged <65 years and16.12/100 000 for those aged ≥65 years, respectively.The older age group had a higher crude mortality forsuicide with psychiatric history than the younger group(14.31/100 000 vs 1.19/100 000), but the opposite wastrue for suicide without psychiatric history (1.81/100 000

vs 3.10/100 000). Elderly women had a higher crudemortality for suicide than elderly men (21.88/100 000 vs10.31/100 000).The overall age-standardised mortality for suicide

declined from 14.22/100 000 in 2003 to 8.39/100 000 in2013 in men, and from 16.82/100 000 to 6.74/100 000in women. In comparison with people aged <65 years,those aged ≥65 years, of both genders, had a highersuicide mortality, especially those with psychiatric history(table 1).The PC for suicide mortality during 2003–2013 was

−37.6% for men and −50.5% for women (tables 2 and3). The suicide mortality decreased more rapidly in menaged ≥65 years than men aged <65 years. The APC wasstatistically significant for decreasing suicide mortalityfor women aged ≥65 years during the period from 2003to 2011 and for men aged <65 years during the periodfrom 2003 to 2008 (figures 1–3).The gender disparity in suicide mortality was greater

in the elderly than in younger people. The male/femaleratio in those aged ≥65 years was 0.5 (χ2=126.7,p<0.001) for overall suicide mortality and 0.4 (χ2=126.9,p<0.001) for suicide mortality with a psychiatric history,while the corresponding ratios in those aged <65 yearswere 1.2 (χ2=2.1, p>0.05) and 1.1 (χ2=0.1, p>0.05),respectively. The mortality rate for suicide without ahistory of psychiatric illness overall showed no genderdifference (M/F ratio 1.01 (χ2=0.013, p>0.05)).The age-specific suicide mortality without a psychiatric

history showed no substantial change with age, whereasthe suicide mortality with psychiatric history increasedafter 65 years of age (figure 4).

DISCUSSIONOur study observed a decreasing suicide mortality rateduring the period from 2003 to 2013 in a rural–urbanfringe district in Shanghai. Elderly people comprisedthe greatest proportion of suicide deaths in Jiading dis-trict. Suicide decedents in the elderly were mostlyfemale, especially for suicide deaths in those peoplewith a psychiatric history. Suicide with psychiatric historymostly occurred in people aged ≥65 years, while suicidewithout psychiatric history mostly in people aged<65 years.Suicide is a serious public health problem, occurring

at all ages. Suicide is the second leading cause of deathin youth and young adults globally, whereas suicide

Table 2 PC and APC for suicide mortality in Jiading, 2003–2013

Male Female

Total <65 ≥65 Total <65 ≥65

PC (%) −37.6 −17.2 −51.5 −50.5 −55.2 −47.5APC (%) −4.9, −0.4, 2.0 −4.8*, 66.8 −29.1, 4.6, −9.7 −2.0, 0.8, −0.5 −26.7*, 25.7, −16.7 −12.2*, −5.2**The APC is significantly different from 0 at α=0.05.APC, annual percentage change; PC, percentage change.

Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227 3

Open Access

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from

mortality is the highest in people aged ≥70 years forboth genders in almost all regions of the world.1 8 9 11 12

Similar findings have also been documented in China.Studies of Suzhou city,13 Chengdu city,14 Harbin prov-ince15 and the whole nation6 all indicated an increase insuicide mortality with age in adults. One study fromShandong province noticed that suicide risk for peopleaged ≥65 years was five times greater than in thegeneral population.16 Previous studies in Shanghai17

have also shown a similar trend with a slightly lowermortality rate of 15.99/100 000 in the elderly than thatof Jiading district (16.12/100 000) in our study.Reasons are complicated. Suicide ideation and suicideattempts in older adults were strongly associated withdisability, activity limitation and extreme pain.18 Otherreasons include the presence of physical or mental dis-orders,19 20 poor function of self-care,21 lower levels ofeducation,22 higher rate of unemployment23 and livingalone.24

The gender difference in suicide mortality in Chinahas changed over the past few decades. One study, basedon data from the China vital registration system for theperiod 1995–1999, found that the mortality rate washigher in women than in men, mainly due to death inyoung rural women.25 Another study, based on datafrom the Chinese Health Statistics Year Books of 2002–2011, noted that the suicide mortality rate in menexceeded that in women after 2006, and proposed thatthe change in women’s social position might be areason.6 Being a rural–urban fringe district, Jiading dis-trict has a male predominance in people aged <65 yearsand a female predominance in people aged ≥65 years.This gender difference in the elderly might be due todifferent status of social support or social connectedness,which is associated with suicidal ideation and suicide inlater life.26 Research has suggested that social factors,such as negative life events and social support, mighthave more significant roles in suicidality in Asia than theWest.27 Another study, conducted in Jiading district,showed a lower score on an objective support subscale,subjective support scale, utility degree of social supportand the total scores of social support in Jiading com-pared with the national norm, and a significantly lowersocial support score in women than men.28 Thesefactors might underlie the predominance of suicide inelderly women in Jiading. Moreover, the difference insuicide method29 or the strain theory30 may explain thegender difference in younger people.Among all the suicide deaths in Jiading, 74.7% were

in people with a psychiatric history, which was higherthan in a previous study in China (62.9%).31 The corre-sponding proportion was even higher in the elderly(88.7%). Having a mental disorder was the most import-ant determinant of suicidality in older adults.22

Insomnia,32 major depressive episodes,33 dysthymia34

and many other disorders are associated with thought ofsuicide or attempts. With ageing of the population, thenumber of elderly people with mental disorders has

Table

3PC

andAPC

forsuicidewithout/withpsychiatric

history

inJiading,2003–2013

Suicidewithoutpsychiatric

history

Suicidewithpsychiatric

history

Male

Female

Male

Female

Total

<65

≥65

Total

<65

≥65

Total

<65

≥65

Total

<65

≥65

PC

(%)

−29.7

−21.1

−57.1

−37.5

−44.1

−12.1

−49.2

−7.5

−50.5

−56.8

−79.6

−51.8

APC

(%)

−0.6,0.01,−0.4

−0.6,−17.3,64.7

−5.5,−13.4

−0.8,0.3

−21.8*,7.6

−38.8,−12.8,48.6,−2.6

−0.4,0.8

−23.8*,23.0

−45.8,−4.9

−1.2,−0.4

−34.0,54.9,−46.6

−9.1*,−0.1

*TheAPC

issignificantlydifferentfrom

0atα=0.05.

APC,annualpercentagechange;PC,percentagechange.

4 Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227

Open Access

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from

increased significantly. Being in a rapid process ofurbanisation, Jiading district might have more emptynesters and elderly people living alone without connec-tion with family. Family connectedness is a strong pro-tective factor against suicide ideation in the elderly, anda key component in the intervention and treatment ofmental illness.35 However, ageing does not appear to

modify the relationship between depressive symptomsand suicide behaviour.36 Distinctive prevention strat-egies should be developed and adopted in two agegroups,37 such as interventions to avoid loneliness andlack of social support status in the elderly38 and treat-ment of mental health conditions in the generalpopulation.39

Figure 2 Joinpoint regression analysis of suicide mortality for men and women without psychiatric history in Jiading, 2003–

2013.

Figure 1 Joinpoint regression analysis of suicide mortality for men and women in Jiading, 2003–2013.

Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227 5

Open Access

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from

LimitationsOur study has several limitations. We did not have theinformation on the suicide attempts of this populationand could not compare the characteristics of suicideand suicide-attempt groups. Second, since the informa-tion on psychiatric disorders of suicide decedents wasprovided by their relatives, this might have led to anunderestimation of this comorbidity. In addition, weassumed that suicide deaths were classified accurately.However, there is evidence that other forms of death

might have been misclassified as suicide, although thisoccurs infrequently.

CONCLUSIONSuicide mortality in the elderly varied and showed adeclining trend during the period 2003–2013 in Jiadingdistrict. The trend was more pronounced in the elderlyfor those with a psychiatric history than for thosewithout such a history. However, the mortality rateremained relatively high for the study period.

Figure 4 Age-specific mortality

(per 100 000) for suicide and for

suicide without/with psychiatric

history in Jiading, 2003–2013.

Figure 3 Joinpoint regression analysis of suicide mortality for men and women with psychiatric history in Jiading, 2003–2013.

6 Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227

Open Access

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from

Author affiliations1Jiading District Center for Disease Control and Prevention, Shanghai, China2Key Laboratory of Public Health Safety of Ministry of Education, School ofPublic Health, Fudan University, Shanghai, China3Faculty of Medicine, School of Epidemiology, Public Health, and PreventiveMedicine, University of Ottawa, Ottawa, Ontario, Canada

Contributors CZ, YS and NW contributed to the study concept and design; allauthors contributed to data acquisition and collection; CZ, YS, NW and YCcontributed to data analysis and interpretation; CZ, NW and YC drafted themanuscript; all authors contributed to revision of the manuscript, read andapproved the final manuscript.

Funding The study was supported by the Shanghai Key DisciplineConstruction Project of Shanghai Municipal Public Health (grant no.15GWZK0801).

Competing interests None declared.

Provenance and peer review Not commissioned; externally peer reviewed.

Data sharing statement No additional data are available.

Open Access This is an Open Access article distributed in accordance withthe Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license,which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, providedthe original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/

REFERENCES1. WHO. Preventing suicide: a global imperative. World Health

Organization, 2014.2. Cai Y, Hu N, Liu Y, et al. Suicide analysis of Chinese population in

2010. Zhonghua Yu Fang Yi Xue Za Zhi 2012;13:480–2.3. Conwell Y, Van Orden K, Caine ED. Suicide in older adults.

Psychiatr Clin North Am 2011;34:451–68.4. Hendin H, Phillips MR, Vijayakumar L, et al. Suicide and suicide

prevention in Asia. Mental health and substance abuse. WorldHealth Organization, 2008.

5. Yip PSF, Liu KY, Hu J, et al. Suicide rates in China during a decadeof rapid social changes. Soc Psychiatry Psychiatr Epidemiol2005;40:792–8.

6. Zhang J, Sun L, Liu Y, et al. The change in suicide rates between2002 and 2011 in China. Suicide Life Threat Behav 2014;44:560–8.

7. ICD-10-CM Tabular list of diseases and injuries 2016. http://apps.who.int/classifications/icd10/browse/2016/en

8. Vichi M, Masocco M, Pompili M, et al. Suicide mortality in Italy from1980 to 2002. Psychiatry Res 2010;175:89–97.

9. Bando DH, Brunoni AR, Fernandes TG, et al. Suicide rates andtrends in São Paulo, Brazil, according to gender, age anddemographic aspects: a joinpoint regression analysis. Rev BrasPsiquiatr 2012;34:286–93.

10. Doll R, Payne P, Waterhouse J. Cancer incidence in five continents.Springer-Verlag Berlin Heidelberg, 1966.

11. Qi X, Hu W, Page A, et al. Dynamic pattern of suicide in Australia,1986–2005: a descriptive-analytic study. BMJ Open 2014;4:e005311–e11.

12. Lovisi GM, Santos SA, Legay L, et al. Epidemiological analysis ofsuicide in Brazil from 1980 to 2006. Rev Bras Psiquiatr 2009;31(Suppl 2):S86-94.

13. Lu Y, Wang L. Analysis on the epidemiological charateristics of thesuicide data of residents in Suzhou from 2003 to 2012. Chin J PrevControl Chron Dis 2014;22:448–50.

14. Peng Z, Feng X, Liao J, et al. Suicide of residents in Chengdu,2006–2010. J Prev Med Info 2013;29:383–5.

15. Lan L, Sui C, Xu G, et al. Analysis on suicide mortality among city’spopulation in Harbin from 1988 to 2009. Chin J Public Health Manag2011(02):196–7.

16. Sun J, Guo X, Zhang J, et al. Suicide rates in Shandong, China,1991–2010: rapid decrease in rural rates and steady increase inmale–female ratio. J Affect Disord 2013;146:361–8.

17. Wei X, Yu Z, Yisheng W. Epidemiological analysis on suicide amongZhabei residents, 1995–2004. Chin J Dis Control Prev2006;10:48–50.

18. Kim SH. Suicidal ideation and suicide attempts in older adults:influences of chronic illness, functional limitations, and pain.Geriatr Nurs 2016;37:9–12.

19. Erlangsen A, Stenager E, Conwell Y. Physical diseases aspredictors of suicide in older adults: a nationwide, register-basedcohort study. Soc Psychiatry Psychiatr Epidemiol 2015;50:1427–39.

20. Ruckenbauer G, Yazdani F, Ravaglia G. Suicide in old age: illnessor autonomous decision of the will? Arch Gerontol Geriatr2007;44:355–8.

21. Tsoh J, Chiu HFK, Duberstein PR, et al. Attempted suicide in elderlyChinese persons: a multi-group, controlled study. Am J GeriatrPsychiatry 2005;13:562–71.

22. Ngamini Ngui A, Vasiliadis HM, Préville M. Individual and area-levelfactors correlated with death by suicide in older adults. Prev Med2015;75:44–8.

23. Wiktorsson S, Runeson B, Skoog I, et al. Attempted suicide in theelderly: characteristics of suicide attempters 70 years and older anda general population comparison group. Am J Geriatr Psychiatry2010;18:57–67.

24. Purcell B, Heisel MJ, Speice J, et al. Family connectednessmoderates the association between living alone. Am J GeriatrPsychiatry 2012;20:717–23.

25. Phillips MR, Li X, Zhang Y. Suicide rates in China, 1995–99. Lancet2002;359:835–40.

26. Fässberg MM, Orden KAV, Duberstein P, et al. A systematicreview of social factors and suicidal behavior in older adulthood.Int J Environ Res Public Health 2012;9:722–45.

27. Park S, Hatim Sulaiman A, Srisurapanont M, et al. The associationof suicide risk with negative life events and social support accordingto gender in Asian patients with major depressive disorder.Psychiatry Res 2015;228:277–82.

28. Zhao J, Zhao X, Liu H, et al. Study on social support statusand influencing factors among community-based elderlyresidents in Jiading, Shanghai. Chin J Health Educ2015;31:11–14, 27.

29. Callanan VJ, Davis MS. Gender differences in suicide methods.Soc Psychiatry Psychiatr Epidemiol 2012;47:857–69.

30. Zhang J, Wieczorek WF, Conwell Y, et al. Psychological strainsand youth suicide in rural China. Soc Sci Med 2011;72:2003–10.

31. Yang GH, Phililips MR, Zhou MG, et al. Understanding the uniquecharacteristics of suicide in China: national psychological autopsystudy. Biomed Environ Sci 2005;18:379.

32. Hung GC-L, Kwok C-L, Yip PSF, et al. Predicting suicide in olderadults—a community-based cohort study in Taipei City, Taiwan.J Affect Disord 2015;172:165–70.

33. Miret M, Ayuso-Mateos JL, Sanchez-Moreno J, et al. Depressivedisorders and suicide: epidemiology, risk factors, and burden.Neurosci Biobehav Rev 2013;37:2372–4.

34. Bernal M, Haro JM, Bernert S, et al. Risk factors for suicidality inEurope: results from the ESEMED study. J Affect Disord2007;101:27–34.

35. Lask J. Family and psychiatric illness: family therapy. In: Wright J, ed.International encyclopedia of the social & behavioral sciences(second edition). Oxford: Elsevier, 2015:745–50.

36. Harvey PD, Czaja SJ, Loewenstein DA. Schizophrenia in later life.Am J Geriatr Psychiatry 2012;20:1–4.

37. Hawton K, van Heeringen K. Suicide. Lancet 2009;373:1372–81.38. Innamorati M, Pompili M, Di Vittorio C, et al. Suicide in the old

elderly: results from one Italian county. Am J Geriatr Psychiatry2014;22:1158–67.

39. Raposo S, El-Gabalawy R, Erickson J, et al. Associations betweenanxiety disorders, suicide ideation, and age in nationallyrepresentative samples of Canadian and American adults. J AnxietyDisord 2014;28:823–9.

Shao Y, et al. BMJ Open 2016;6:e012227. doi:10.1136/bmjopen-2016-012227 7

Open Access

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://bmjopen.bm

j.com/

BM

J Open: first published as 10.1136/bm

jopen-2016-012227 on 19 August 2016. D

ownloaded from